Sizing hydrogen stations accurately is a very important aspect of planning for a successful hydrogen refueling network. Over-sizing a station results in wasted capital, under-sizing a station results in unsatisfied customers. Determining the important metrics to get the correct sizing is important in minimizing these risks. To analyze this issue, we started with the hypothesis that hydrogen demand distribution may be similar to gasoline demand distribution in some respects. Evidence suggests that many people refuel near either work or home[1]. The extent to which this is true is investigated. The number of gasoline stations is great enough so that consumers can refuel where they would like. By understanding their behavior, we can better understand what demand should be expected at a hydrogen station. In this analysis we assume that fuel demand distribution correlates to either population factors and/or traffic factors. Factors to be tested include: total traffic density by time period, pass-by traffic by time period, population, income, rush hour origins, rush hour destinations, and car ownership. The extent to which each of these factors influence the number of gallons pumped at a station will be analyzed by a regression analysis. Preliminary results show that on a large scale, population is a better predictor of fuel demand. However, as the analysis area is divided into smaller and smaller zones, traffic becomes a better predictor (see Figure1). Some interesting patterns also emerge. Customers seem less likely to refuel in the central business district than traffic patterns may indicate. This may be an important factor in sizing a station. Also, customers appear more likely to refuel between home and the freeway as opposed to any other part of his/her journey.

1 Kitamura, Ryuchi. and Dan Sperling. 1987. Refueling Behavior of Automobile Drivers. Transportation Research 21A, no. No. 3: 235-245.